Post on 28-Apr-2023
Hindawi Publishing CorporationInternational Journal of Cell BiologyVolume 2013, Article ID 297537, 15 pageshttp://dx.doi.org/10.1155/2013/297537
Research ArticleModulation of Vascular Cell Function by Bim Expression
Margaret E. Morrison,1 Tammy L. Palenski,2 Nasim Jamali,2
Nader Sheibani,2,3 and Christine M. Sorenson1,3
1 Department of Pediatrics, University of Wisconsin School of Medicine and Public Health, 600 Highland Avenue, H4/444 CSC,Madison, WI 53792-4108, USA
2Department of Ophthalmology and Visual Sciences, University of Wisconsin School of Medicine and Public Health, Madison,WI 53792, USA
3McPherson Eye Research Institute, University of Wisconsin School of Medicine and Public Health, Madison, WI 53792, USA
Correspondence should be addressed to Christine M. Sorenson; cmsorenson@pediatrics.wisc.edu
Received 24 July 2013; Accepted 4 September 2013
Academic Editor: Claudia Giampietri
Copyright © 2013 Margaret E. Morrison et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.
Apoptosis of vascular cells, including pericytes and endothelial cells, contributes to disease pathogenesis in which vascularrarefaction plays a central role. Bim is a proapoptotic protein that modulates not only apoptosis but also cellular functions such asmigration and extracellular matrix (ECM) protein expression. Endothelial cells and pericytes each make a unique contribution tovascular formation and function although the details require further delineation. Here we set out to determine the cell autonomousimpact of Bim expression on retinal endothelial cell and pericyte function using cells prepared from Bim deficient (Bim−/−)mice. Bim−/− endothelial cells displayed an increased production of ECM proteins, proliferation, migration, adhesion, and VEGFexpression but, a decreased eNOS expression and nitric oxide production. In contrast, pericyte proliferation decreased in theabsence of Bimwhilemigration, adhesion, andVEGF expression were increased. In addition, we demonstrated that the coculturingof either wild-type or Bim−/− endothelial cells with Bim−/− pericytes diminished their capillary morphogenesis. Thus, our datafurther emphasizes the importance of vascular cell autonomous regulatory mechanisms in modulation of vascular function.
1. Background
Apoptosis facilitates the removal of unwanted cells duringdevelopment and maintains tissue homeostasis. Bcl-2 familymembers influence apoptosis in either a positive or a negativefashion. Family members are classically grouped into threesubclasses including one that inhibits apoptosis, a secondthat induces apoptosis and a third contains that familymembers, such as Bim, that only have a BH3 domain thatbinds antiapoptotic family members to promote apoptosis[1]. Bcl-2 exerts opposing effects with regards to apoptosiscompared with Bim, consistent with their opposing effectson cell adhesion and migration [2]. The removal of a singleallele of Bim is sufficient to prevent the degenerative disorderscaused by Bcl-2 deficiency [3, 4]. Bim expression is alsoessential for apoptosis of a wide range of growth factordeprived cells [5], including endothelial cells and pericytes
[6]. Therefore, improper modulation of apoptosis couldimpact the development and/or the pathogenesis of manydiseases including diabetic retinopathy.
Murine retinal vascular development proceeds after birthand is complete by postnatal day 21 (P21). Remodeling andpruning of the retinal vasculature continue until P42 [7, 8].Pro- and anti-apoptotic factors regulate apoptosis duringretinal vascular development, remodeling, and regression[8–10]. Bim influences not only apoptosis but also celladhesion, migration, and extracellular matrix (ECM) proteinexpression. The ability of Bim to impact both apoptosis andthe ECM milieu is integral to its role during retinal vesselremodeling, and regression. However, whether Bim expres-sion impacts retinal endothelial cell and pericyte function ina similar manner remains to be determined.
Our previous studies demonstrated that Bim modulatesretinal vascular development, remodeling, and regression
2 International Journal of Cell Biology
[11]. Bim deficient (Bim−/−) mice demonstrated precociousformation of the retinal vascular plexus, increased vasculardensity, and attenuated hyaloid vessel regression. Loss ofBim expression also protected the retinal vasculature fromhyperoxia-mediated vessel obliteration and neovasculariza-tion during oxygen induced ischemic retinopathy (OIR) [11].These studies indicated that Bim expression is essential forretinal vascular remodeling but did not implicate a specificvascular cell type. Thus, understanding the role Bim plays inmodulating endothelial cell and pericyte function will give usfurther insight into how aberrant regulation of retinal vesseldevelopment and remodeling can be avoided to preservevision.
Endothelial cells and pericytes play specialized roles inthe retina during the development and maintenance of thevasculature. Pericyte coverage of vascular sprouts stabilizesthe vessels while endothelial cells provide the inner liningof the blood vessel [12, 13]. Retinal endothelial cell andpericyte dysfunction can lead to retinal vascular rarefac-tion and subsequent vision loss as the one that occurs indiabetic retinopathy. Here, we assessed whether the lossof the proapoptotic protein Bim differentially impacts theretinal endothelial cell and pericyte function. We observedan increased VEGF expression, ECM protein production,cell migration, proliferation, and adhesion in Bim−/− retinalendothelial cells compared with wild-type cells. Althoughthe decreased endothelial nitric oxide synthase (eNOS)expression corresponded with decreased nitric oxide (NO)production in Bim−/− retinal endothelial cells, capillary mor-phogenesis was similar to that observed in wild-type retinalendothelial cells. In contrast to Bim−/− retinal endothelialcells, Bim−/− pericytes displayed decreased proliferationcompared with their wild-type counterpart. Bim−/− peri-cytes also demonstrated increased migration, adhesion, andVEGF expression. Coculturing wild-type or Bim−/− retinalendothelial cells with Bim−/− pericytes diminished capillarymorphogenesis.Thedata presented here demonstrate that thedeletion of Bim differentially impacts retinal endothelial celland pericyte function, perhaps through the modulation ofresponse to their microenvironment.
2. Results
2.1. Bim−/− Retinal Endothelial Cells and Pericytes Demon-strate Aberrant Proliferation. To further investigate the roleBim plays in retinal vascular function, we isolated retinalendothelial cells and pericytes from wild-type and Bim−/−mice as previously described [2, 14–17]. We first examinedretinal endothelial cell and pericyte morphology and theirexpression of cell specific markers to confirm that these cellsmaintain their endothelial cell and pericyte characteristics.Bim−/− retinal endothelial cells displayed a slightly elongatedmorphology compared to wild-type cells when plated ongelatin-coated plates (Figure 1(a)). The morphology of wild-type and Bim−/− retinal pericytes was similar (Figure 1(b)).Wild-type and Bim−/− retinal endothelial cells expressedthe endothelial cell markers VE-cadherin and PECAM-1
(Figure 1(c)). Retinal pericytes expressed the pericyte mark-ers platelet derived growth factor-receptor 𝛽 (PDGFR𝛽) andneuroglia proteoglycan 2 (NG2) (Figure 1(d)).
Next we examined the rate of retinal endothelial celland pericyte apoptosis and proliferation. Wild-type andBim−/− retinal endothelial cells and pericytes were incubatedwith staurosporine to induce apoptosis. Although the basallevels of apoptosis were similar, upon incubation with stau-rosporine for 24 h, Bim−/− endothelial cells and pericytesdemonstrated less apoptosis compared to wild-type cells(Figures 2(a) and 2(b)). Next cell proliferation was examinedby counting the number of cells every other day for 2weeks. Figure 2(c) shows that Bim−/− retinal endothelial cellsproliferated at a faster rate than their wild-type counterpart.In contrast, Bim−/− pericytes demonstrated a reduced rateof proliferation compared to their wild-type counterpart(Figure 2(d)).
2.2. Loss of Bim Expression Enhances Retinal Endothelial Celland Pericyte Migration. An appropriate rate of migration isessential for optimal capillarymorphogenesis. Herewe exam-ined cell migration characteristics using scratch wound andtranswell migration assays. A confluent monolayer of wild-type and Bim−/− endothelial cells (Figure 3(a)) and pericytes(Figure 3(b)) was wounded, and 5-fluorouracil (5-FU) wasadded to prevent cell proliferation. Both Bim deficient retinalendothelial cells and pericytes migrated faster than theirwild-type counterparts (Figures 3(a) and 3(b)). A quantitativeassessment of the scratch wound data for endothelial cellsand pericytes is shown in Figures 3(c) and 3(d), respectively.Enhancedmigration of Bim−/− endothelial cells and pericyteswas also observed using a transwell migration assay as shownin Figures 3(e) and 3(f), respectively. Thus, the lack of Bimexpression enhances the migration of retinal vascular cells.
2.3. Bim−/− Retinal Vascular Cells Were More Adherent.Changes in migration of Bim−/− retinal vascular cells couldbe due to altered cell adhesion. We next examined the abilityof wild-type and Bim−/− endothelial cells (Figure 4(a)) andpericytes (Figure 4(b)) to adhere to fibronectin, collagen I,collagen IV, and vitronectin. Bim deficient endothelial cellsand pericytes displayed increased adhesion to all of thesematrices compared to their wild-type counterpart.
Changes in cell migration and adhesion could be theresult of aberrant integrin expression. Next we analyzed theintegrin expression on the surface of retinal endothelial cellsand pericytes by FACScan analysis (Figures 5(a) and 5(b)).Wild-type and Bim−/− endothelial cells expressed similarlevels of 𝛼5, 𝛼v𝛽3, and 𝛽1 integrins on their surface whilewild-type and Bim−/− retinal pericytes expressed similarlevels of 𝛼2, 𝛼v𝛽3, 𝛽1, and 𝛽3 integrins. Thus, the increasedadhesion noted in Bim−/− retinal vascular cells may beindependent of significant changes in the integrin expressionand may be due to alterations in the affinity and/or avidity ofthese integrins.
International Journal of Cell Biology 3
WTEn
doth
elia
l cel
lsBim−/−
(a)
WT
Peric
ytes
Bim−/−
(b)
IgGPECAM-1
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
WT
IgGPECAM-1
FL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
IgGVE-cadherin
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
IgGVE-cadherin
FL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
Bim−/−
(c)
IgGFL-1 FITC
Max
(%)
100
80
60
40
20
0104103102101100
WT
IgG
FL-1 FITC
Max
(%)
100
80
60
40
20
0104103102101100
IgGNG2
FL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
IgGNG2
FL-1 FITC
100
80
60
40
20
0M
ax (%
)104103102101100
PDGFR-𝛽 PDGFR-𝛽
Bim−/−
(d)
Figure 1: Wild-type and Bim−/− cells morphology. Wild-type and Bim−/− retinal endothelial cells (panel (a)) and pericytes (panel (b)) werecultured on gelatin or uncoated plates, respectively. Cells were photographed using a phase microscope in digital format at lowmagnification(×40). In panel (c), retinal endothelial cells prepared from wild-type and Bim−/− mice were examined for the expression of PECAM-1 andVE-cadherin FACScan analysis. In panel (d), retinal pericytes from wild-type and Bim−/− mice were examined for the expression of NG2 andPDGFR-𝛽 by FACScan analysis. The shaded areas show staining in the presence of control IgG.
2.4. Increased Tenascin C and Osteopontin Expression inBim−/− Retinal Endothelial Cells. Modulating the ECMmilieu can influence vascular cell functions including celladhesion andmigration. To examine whether the lack of Bimexpression differentially impacts retinal endothelial cell orpericyte ECM expression, serum-free conditioned mediumfrom these retinal vascular cells was evaluated by Westernblot analysis. Retinal endothelial cells from Bim−/− micedisplayed an increased expression of tenascin C, osteopontin,fibronectin, and TSP1 compared to their wild-type coun-terpart (Figure 6(a)). In contrast, Bim−/− retinal pericytesdisplayed similar levels of tenascin C, fibronectin, but an
increased TSP1 and a decreased osteopontin expression(Figure 6(b)).
2.5. Decreased p-eNOS Expression in Bim−/− Retinal Endothe-lial Cells. The activation of eNOS and Akt1 by VEGF pro-motes angiogenesis [18–20]. In the absence of Bim, VEGFexpression was dramatically increased in retinal endothelialcells and pericytes (Figures 7(a) and 7(b)). However, inBim−/− retinal endothelial cells, the increased VEGF expres-sion did not affect Akt, phospho-Akt, or HSP90 expression(Figure 7(c)). Bim−/− retinal endothelial cells also demon-strated nearly undetectable eNOS expression (Figure 7(c))
4 International Journal of Cell Biology
300000
250000
200000
150000
100000
50000
0
Mea
n lu
min
esce
nce (
RLU
)
Control StaurosporineWT WT
∗∗∗
Bim−/− Bim−/−
(a)
300000
250000
200000
150000
100000
50000
0
Mea
n lu
min
esce
nce (
RLU
)
Control StaurosporineWT WT
∗∗∗
Bim−/− Bim−/−
(b)
300250200150100
500
Day2 4 6 8 10 12
Cel
l num
ber (×10
−4)
(c)
250
200
150
100
50
0
Day2 4 6 8 10 12
Cel
l num
ber (×10
−4)
(d)
Figure 2:Modulation of proliferation in Bim−/− cells.Wild-type and Bim−/− retinal endothelial cells (panel (a)) and pericytes (panel (b)) wereincubated with 10 nM staurosporine for 24 hours. Apoptosis was determined using a Caspase-Glo 3/7 assay. Wild-type and Bim−/− retinalendothelial cells (panel (c)) and pericytes (panel (d)) were monitored for their growth rate over a two-week time frame for wild-type (I) andBim−/− (◻) cells. These experiments were repeated twice with similar results (∗∗∗𝑃 < 0.001).
which corresponded with a 25-fold decrease in NO produc-tion (Figure 7(d)). Thus, VEGF activation of eNOS throughthe activation of Akt is uncoupled in the absence of Bim.
2.6. Bim−/− Pericytes Disrupt Retinal Endothelial Cell Capil-lary Morphogenesis. Capillary morphogenesis is fundamen-tal in vascular development and remodeling in which peri-cytes play a major role.The ability of pericytes to migrate andproduce survival factors is central to their recruitment alongthe developing endothelium. However, pericytes typically donot undergo any substantial morphogenesis in Matrigel [21].In the retina, endothelial cells and pericytes function togetherto facilitate retinal vascularization. Here, we evaluated theability of pericytes fromwild-type and Bim−/−mice to impactretinal endothelial cell capillary morphogenesis.
Wild-type and Bim−/− endothelial cells demonstrateda similar number of branch points in Matrigel (Figures8(a)–8(c)). However, Bim−/− endothelial cells formed ratherstunted capillary-like tubes compared with wild-type cells(Figures 8(a)–8(c)). Next, we cocultured wild-type retinalendothelial cells with either wild-type or Bim−/− pericytes.Wild-type retinal endothelial cells cocultured with wild-typepericytes demonstrated an improved tubular network com-pared to wild-type endothelial cells alone (Figures 8(a)–8(c)).
In contrast, wild-type retinal endothelial cells coculturedwithBim−/− pericytes demonstrated diminished capillary mor-phogenesis. Bim−/− endothelial cells cocultured with wild-type pericytes demonstrated enhanced capillarymorphogen-esis while cocultures of Bim−/− endothelial cells and pericytesdemonstrated diminished capillary morphogenesis (Figures8(a) and 8(c)). Thus, coculturing of retinal endothelial cellswith Bim−/− pericytes resulted in diminished endothelial cellcapillary morphogenesis.
Bim−/− pericytes demonstrated increased VEGF levelscompared to their wild-type counterpart. Next, we addressedwhether an increased VEGF expression contributed to theinability of wild-type endothelial cells to undergo capillarymorphogenesis when cocultured with Bim−/− pericytes.VEGFR-1 (Flt-1) has a higher affinity for VEGF compared toVEGFR-2 and can compete for VEGF binding to VEGFR-2and its activation. Thus, the soluble form of Flt-1 can actas a trap for VEGF and prevent its signaling. Here, weused sFlt-1 FC chimera to demonstrate that the negativeimpact of Bim−/− pericytes on capillary morphogenesis ofretinal endothelial cells is driven by the production of excessVEGF. The sFlt1 was added to wild-type endothelial celland Bim−/− pericyte coculture experiments, and capillarymorphogenesis was monitored (Figure 8(d)). As shown
International Journal of Cell Biology 5
WT
0 hr
48 hr
Bim−/−
(a)
WT
72 hr
0 hr
Bim−/−
(b)
150
100
50
0
Wou
nd cl
osur
e (%
)
WT
∗∗∗
Bim−/−
(c)
150
100
50
0
Wou
nd cl
osur
e (%
)
WT
∗∗∗
Bim−/−
(d)
WT
∗∗∗
25
20
15
10
5
0
Mea
n nu
mbe
r of c
ells/
field
Bim−/−
(e)
WT
∗∗∗
20
15
10
5
0
Mea
n nu
mbe
r of c
ells/
field
Bim−/−
(f)
Figure 3: Bim−/− retinal endothelial cells and pericytes demonstrate increased migration. Cell migration was determined by the scratchwounding of retinal endothelial cells (panel (a)) or pericytes (panel (b)) monolayers, and wound closure was monitored using a phasemicroscope in digital format at low magnification (×40). A representative experiment is shown here. Please note that wild-type vascular cellsmigrate slower than Bim−/− cells. The quantitative assessment of the data is shown in panels (c) and (d). Transwell assays were performedwith wild-type and Bim−/− retinal endothelial cells (panel (e)) and pericytes (panel (f)). These experiments were repeated twice with similarresults (∗∗∗𝑃 < 0.001).
6 International Journal of Cell Biology
20 4 6 8 10
2.5
2.0
1.5
1.0
0.5
0.0
Collagen I (𝜇g/mL)
OD405
nm
20 4 6 8 10
2.5
2.0
1.5
1.0
0.5
0.0
Collagen IV (𝜇g/mL)
OD405
nm
20 4 6 8 10
2.5
2.0
1.5
1.0
0.5
0.0
Fibronectin (𝜇g/mL)
WT
OD405
nm
20 4 6 8 10
2.5
2.0
1.5
1.0
0.5
0.0
Vitronectin (𝜇g/mL)
WT
OD405
nm
Bim−/− Bim−/−
(a)
20 4 6 8 10
3
2
1
0
Collagen I (𝜇g/mL)
OD405
nm
20 4 6 8 10
3
2
1
0
Collagen IV (𝜇g/mL)
OD405
nm
20 4 6 8 10
3
2
1
0
Vitronectin (𝜇g/mL)WT
OD405
nm
20 4 6 8 10
3
2
1
0
Fibronectin (𝜇g/mL)
WT
OD405
nm
Bim−/−Bim−/−
(b)
Figure 4: Bim−/− endothelial cells and pericytes were more adherent. The adhesion of retinal wild-type (◼) and Bim−/− (I) endothelial cells(panel (a)) and pericytes (panel (b)) to fibronectin, vitronectin, collagen type I, or collagen IV was determined as described in Section 5.Please note that Bim−/− vascular cells had increased adherence. These experiments were repeated twice with similar results.
International Journal of Cell Biology 7
IgG
100
80
60
40
20
0
Max
(%)
104103102101100
FL-1 FITC
IgG
FL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
IgG
100
80
60
40
20
0
Max
(%)
FL-1 FITC104103102101100
IgG
100
80
60
40
20
0
Max
(%)
FL-1 FITC104103102101100
IgG
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
WT
IgGFL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
𝛽1 𝛽1
𝛼5 𝛼5
𝛼v𝛽3 𝛼v𝛽3
Bim−/−
(a)
IgG
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
WT
IgG
FL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
IgG
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
IgG
FL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
IgG
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
IgGFL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
IgG
100
FL-1 FITC
80
60
40
20
0
Max
(%)
104103102101100
IgGFL-1 FITC
100
80
60
40
20
0
Max
(%)
104103102101100
𝛽1
𝛽3
𝛼5
𝛽1
𝛽3
𝛼5
𝛼v𝛽3 𝛼v𝛽3
Bim−/−
(b)
Figure 5: Wild-type and Bim−/− retinal vascular cells had similar integrin expression. Expression of various integrins was determined inwild-type and Bim−/− endothelial cells (panel (a)) and pericytes (panel (b)) using FACScan analysis as described in Section 5. The shadedgraphs show staining in the presence of control IgG. These experiments were repeated twice with similar results.
8 International Journal of Cell Biology
Tenascin C
Osteopontin
Fibronectin
TSP1
WT
𝛽-Actin
Bim−/−
(a)
WT Bim−/−
(b)
Figure 6: Altered expression of ECM proteins in Bim−/− retinalcells. Wild-type and Bim−/− retinal endothelial cells (panel (a))and pericytes (panel (b)) were grown for 2 days in serum-freemedium.Themediumwas harvested, clarified, andWestern-blottedfor extracellular matrix proteins as noted. The expression of 𝛽-actin from total cell lysates was used as a loading control. Theseexperiments were repeated twice with similar results.
above, coculture of wild-type endothelial cells with Bim−/−pericytes demonstrated decreased numbers of branch pointscompared towild-type endothelial cells alone.The addition ofsFlt1 to wild-type endothelial and Bim−/− pericyte cocultureexperiments significantly restored branching morphogenesisin Matrigel. The addition of control IgG to wild-typeendothelial cell and Bim−/− pericyte coculture experimentsdid not impact the number of branch points (Figures 8(d)and 8(e)). Therefore, enhanced VEGF expression contributesto the decreased capillary morphogenesis of endothelial cellscocultured with Bim−/− pericytes.
3. Discussion
Aberrant modulation of apoptosis has a reoccurring themein many diseases that result in blindness. The regressionof the hyaloid vasculature is an apoptosis-driven processin which Bim expression plays a central role [11]. Failureof the hyaloid vasculature to regress is a common congen-ital ocular malformation that can lead to blindness [22].Impropermodulation of apoptosis can also contribute to reti-nal vascular rarefaction and/or neovascularization, observedduring retinopathy of prematurity and diabetic retinopathy.Thus, understanding how apoptosis is modulated in retinalvascular cells, including endothelial cells and pericytes, willfurther our understanding of normal and abnormal retinalvascularization and vascular cell autonomous functions.
The retina has the highest numbers of pericytes cover-ing its blood vessels compared to other tissues [23]. Theappropriate interaction of pericytes with each other andendothelial cells is important for vessel maturation and stabi-lization. Altered pericyte migration or adhesion may disruptor enhance these interactions leading to vasculopathies orincreased vascular stabilization. Here, we show that pericytemigration and adhesion are enhanced in the absence ofBim perhaps as a result of their increased VEGF expression.Increased pericyte migration may lead to greater pericytecoverage of the retinal vasculature, which may contribute tothe precocious formation of the deep vascular plexus andprotection from hyperoxia-mediated vessel obliteration thatwe previously observed [11].
Maturation of the retinal vasculature involves remodelingand pruning of vessels into an organized branched networkof vessels to meet the demand of the tissue. Modulation ofapoptosis by Bcl-2 family members facilitates retinal vascularremodeling and regression [11, 24]. Bim expression is requiredfor retinal capillary pruning, hyaloid vessel regression, andhyperoxia-induced retinal vessel obliteration [11]. Here, wehave extended these findings to delineate the role Bim playsin retinal endothelial cell and pericyte function. We showthat in addition to enhanced migration, retinal pericytesfrom Bim−/− mice were resistant to an apoptotic challengeand demonstrated an increased adhesion to extracellularmatrices. Extracellularmatrix protein expression by pericyteswas only modestly impacted by the loss of Bim expression.Perhaps an increased pericyte adhesion in the absence of Bimenhances pericyte survival and coverage of retinal vessels.This may explain the resistance of the retinal vasculature ofBim−/− mice to developmental pruning or excessive pruningin response to hyperoxia. Therefore, modulating pericyteadhesion and migratory properties may be essential forretinal vascular pruning.
VEGF expression is thought to be a key player in main-taining endothelial cell homeostasis. Since VEGF withdrawalcan lead to endothelial cell apoptosis, determining howVEGFlevels can be modulated by expression of Bcl-2 family mem-bers may aid in our understanding of normal and aberrantretinal vascularization and pruning. Here, we show that Bimdeficiency increases VEGF levels in both endothelial cellsand pericytes. Pericytes act as an angiogenic soup, producing7-fold higher VEGF levels than their endothelial cell coun-terpart (Figure 7). In the absence of Bim, VEGF expressionin retinal endothelial cells increased to nearly twice thatobserved in wild-type pericytes. Unfortunately, VEGF canalso inhibit blood vessel maturation and ablate pericytecoverage of nascent sprouts causing vessel destabilization[25].The increased VEGF produced by Bim−/− pericytes mayresult in the diminished capillary morphogenesis observedwhen cocultured with either wild-type or Bim−/− endothelialcells. Greenberg et al. [25] have also defined VEGF asan inhibitor of neovascularization which is consistent withthe inability of Bim−/− mice to undergo neovascularizationfollowing hyperoxia-induced ischemia [11]. However, retinaldevelopment in these mice proceeds quite well with preco-cious formation of the retinal vascular plexus and increased
International Journal of Cell Biology 9
0.3
0.2
0.1
0.0
60
40
20
0
VEG
F (p
g/m
L)
OD450
∗∗∗
WT Bim−/−
(a)
160
120
80
40
0
VEG
F (p
g/m
L)
0.6
0.4
0.2
0.0
OD450
∗∗∗
WT Bim−/−
(b)
WT
Phospho-eNOS
eNOS
Phospho-Akt
Akt
HSP90
𝛽-Actin
Bim−/−
(c)
150
100
50
0∗∗∗
WT
DA
F-FM
fluo
resc
ence
Bim−/−
(d)
Figure 7: Decreased eNOS and increased VEGF expressions in Bim−/− vascular cells. In panels (a) and (b), an immunoassay was used todetermine VEGF levels (pg/mL) in retinal endothelial cells (a) and pericytes (b) from wild-type and Bim−/− mice. Protein lysates (20𝜇g)from wild-type and Bim−/− retinal endothelial cells were analyzed by Western blot analysis for the expression of phospho-eNOS, eNOS,HSP90, phospho-Akt, and Akt (panel (c)). 𝛽-Actin expression was assessed as a loading control (panel (c)). In panel (d), the intracellular NOproduction was determined as analyzed byDAF-FMfluorescence for retinal endothelial cells. Please note that the increased VEGF expressionwas independent of eNOS expression in Bim−/− endothelial cells (∗∗∗𝑃 < 0.001).
vascular density in contrast to what has been previouslyshown in fibrosarcomas [11, 25]. Thus, the ability of Bim toimpact cell survival may be, in part, through the modulationof VEGF expression.
Phosphorylation of eNOS canmediate the proangiogenicactivity of VEGF. Here, we show an increased VEGF expres-sion but a decreased eNOS expression in Bim−/− retinalendothelial cells, consistent with our previous studies inBim−/− kidney and lung endothelial cells [2]. Moreover,our recent studies further support a reciprocal relationshipbetween VEGF and eNOS expressions in endothelial cells
[26]. Our previous studies demonstrated that knocking downeNOS expression in kidney endothelial cells from diabeticmice not only decreased VEGF expression but also restoredmigration and capillary morphogenesis [26]. Thus, local reg-ulation of Bim expression may play a central role in vascularpruning through modulating eNOS and VEGF expressions.
4. Conclusions
Our previous studies demonstrated an inability of the retinasfrom Bim−/− mice to undergo retinal vascular pruning
10 International Journal of Cell Biology
WT REC + WT PC
WT REC Bim−/− REC
Bim−/− REC + Bim−/− PC
WT REC + Bim−/− PC Bim−/− REC + WT PC
(a)
WT
REC
WT
REC+
WT
PC
50
40
30
20
10
0
Mea
n nu
mbe
r of
bran
ch p
oint
s/fie
ld
∗∗∗∗ ∗∗∗∗
WT
REC
+ Bi
m−/−
PC
(b)
50
40
30
20
10
0
Mea
n nu
mbe
r of
bran
ch p
oint
s/fie
ld
∗∗∗∗ ∗∗∗∗
Bim
−/−
REC
Bim
−/−
REC
+ Bi
m−/−
PC
Bim
−/−
REC
+ W
T PC
(c)
Figure 8: Continued.
International Journal of Cell Biology 11
WT REC WT REC + Bim−/− PC
WT REC + Bim−/− PC + sFlt1 WT REC + Bim−/− PC + IgG
(d)
25
20
15
10
5
0
Mea
n nu
mbe
r of
bran
ch p
oint
s/fie
ld
WT
REC
∗∗∗∗ ∗∗∗∗
WT
REC
+ Bi
m−/−
PC +
sFlt1
WT
REC
+ Bi
m−/−
PC +
IgG
WT
REC
+ Bi
m−/−
PC
(e)
Figure 8: Bim−/− pericytes diminished capillary morphogenesis of retinal endothelial cells. Wild-type (WT) and Bim−/− retinal endothelialcells (REC) were plated alone or with wild-type or Bim−/− pericytes (PC) in Matrigel and photographed in digital format (panel (a)). Thequantitative assessment of the data is shown in panels (b) and (c). In panel (d), wild-type retinal endothelial cells were plated alone or withBim−/− pericytes in the presence or absence of sFlt1 or IgG control (10𝜇g/mL) and photographed. Panel (e) is the quantitative assessment ofthe data in panel (d). The data are the mean number of branch points per field (×40) ± SD. These experiments were repeated at least twicewith similar results (∗∗∗𝑃 < 0.001; ∗∗∗∗𝑃 < 0.0001).
12 International Journal of Cell Biology
Endothelial cells Pericytes
Angiogenesis/remodeling/regression
Bim deficiency
↑ VEGF↑ Proliferation↑ Migration↑ AdhesionAltered ECM expression
↑ VEGF
↑ Migration↑ AdhesionAltered ECM expression
↓ Proliferation
Figure 9: A summary of the modulation of VEGF expression, proliferation, migration, adhesion, and ECM expression in Bim deficientendothelial cells and pericytes.
leading to increased retinal cellularity and capillary loops[11]. These mice were also resistant to hyperoxia-mediatedvessel obliteration and neovascularization during OIR, aswell as a hyaloid vessel regression. Thus, Bim expressionmay modulate retinal vascular pruning by localized changesin retinal vascular cell adhesion, migration, and VEGFexpression (Figure 9) thereby, either protecting or inducingpruning by localized apoptosis.
5. Methods
5.1. Experimental Animals and Cell Cultures. The mice usedfor these studies were maintained and treated in accor-dance with our protocol approved by the University ofWisconsin Animal Care and Use Committee. Immortomiceexpressing a temperature-sensitive SV40 large T antigenwere obtained fromCharles River Laboratories (Wilmington,MA, USA). Bim−/− mice (Jackson Laboratory, Bar Harbor,ME, USA) were crossed with Immortomice and screenedas previously described [2, 17]. To isolate retinal endothe-lial cells, retinas from approximately four- to five-week-old wild type and Bim−/− Immortomice were dissected outaseptically and placed in serum-free Dulbecco’s ModifiedEagle Medium (DMEM) containing penicillin/streptomycin(Sigma, St. Louis, MO, USA). The retinas were pooledtogether, rinsed with DMEM, minced into small pieces in a60mm tissue culture dish using sterilized razor blades, anddigested in 5mL of collagenase type I (1mg/mL in serum-free DMEM; Worthington, Lakewood, NJ, USA) for 30–45minutes at 37∘C. Following digestion, DMEMcontaining 10%fetal bovine serum (FBS) was added and cells were pelleted.The cellular digests were then filtered through a double layerof sterile 40 𝜇m nylon mesh (Sefar America Inc., HanoverPark, IL, USA), centrifuged at 400×g for 10min to pelletcells, and then washed twice with DMEM containing 10%FBS. The cells were resuspended in 1.5mL medium (DMEMwith 10% FBS) and incubated with sheep anti-rat magneticbeads precoated with anti-PECAM-1 antibody (MEC13.3,BD Biosciences, Bedford, MA, USA) as described previously[27]. After affinity binding, magnetic beads were washedsix times with DMEM containing 10% FBS. Bound cells inendothelial cell growth medium were plated into a single
well of a 24-well plate precoated with 2 𝜇g/mL of humanfibronectin (BD Biosciences). Endothelial cells were grownin DMEM containing 10% FBS, 2mM L-glutamine, 2mMsodium pyruvate, 20mM HEPES, 1% nonessential aminoacids, 100 𝜇g/mL streptomycin, 100U/mL penicillin, heparinat 55U/mL (Sigma, St. Louis, MO, USA), endothelial growthsupplement 100𝜇g/mL (Sigma, St. Louis, MO, USA), andmurine recombinant interferon-𝛾 (R&D, Minneapolis, MN,USA) at 44 units/mL. Cells were maintained at 33∘C with5% CO
2
. Cells were progressively passed to larger plates,maintained, and propagated in 1% gelatin-coated 60mmdishes. Retinal endothelial cells were positive for B4-lectin(a mouse endothelial cell specific lectin) and expressedPECAM-1 and VE-cadherin as previously described [2, 14].
Pericytes were isolated from mouse retina by collectingretinas from one litter (6-7 pups, 4 weeks old) using adissecting microscope. Retinas (12 to 14) were rinsed withserum-free Dulbecco’s Modified Eagle’s Medium (Sigma),pooled in a 60mmdish,minced, and digested for 45minwithcollagenase type II (1mg/mL; Worthington, Lakewood, NJ,USA) with 0.1% bovine serum albumin (BSA) in serum-freeDMEM at 37∘C. Cells were rinsed in DMEM containing 10%fetal bovine serum (FBS) and centrifuged for 5min at 400×g.Digested tissue was resuspended in 4mL of DMEM contain-ing 10% FBS, 2mM L-glutamine, 100 𝜇g/mL streptomycin,100U/mLpenicillin, and themurine recombinant interferon-𝛾 (R&D, Minneapolis, MN, USA) at 44U/mL. Cells, tissue,and medium were evenly divided into 4 wells of a 24-welltissue culture plate and maintained at 33∘C with 5% CO
2
.Cells were progressively passed to larger plates, maintained,and propagated in 60mm dishes [21]. The experimentsdescribed here were performed with two separate isolationsof cells with similar results.
5.2. Cell Apoptosis Assays. Apoptosis was determined bymeasuring caspase activation using a Caspase-Glo 3/7-assaykit as recommended by the supplier (Promega, Madison,WI,USA).The assay provides caspase-3/7 DEVD-aminoluciferinsubstrate and the caspase 3/7 activity is detected by lumi-nescent signal. For the assay, cells were plated at 8 × 103 perwell of a 96-well plate. As an oxidative or apoptotic stimulus,cells were incubated with 10 nM staurosporine (Invitrogen),
International Journal of Cell Biology 13
in growth medium for 24 h at 33∘C. Caspase activity wasdetected using a luminescent microplate reader (Victa2 1420Multilabel Counter, PerkinElmer, Waltham, MA, USA). Allsamples were prepared in triplicate and repeated twice withsimilar results [15].
5.3. Scratch Wound Assay. Cells (4 × 105) were plated in60mm tissue culture dishes and allowed to reach conflu-ence (2-3 days). After aspirating the medium, cell layerswere wounded using a 1mL micropipette tip. Plates werethen rinsed with PBS, fed with growth medium containing100 ng/mL of 5-fluorouracil (5-FU), to rule out potentialcontribution of differences in cell proliferation. The woundswere observed and photographed every 24 hours for up to72 hours. The distance migrated was determined as a percentof total distance for quantitative assessments as describedpreviously [2].These experiments were repeated at least twicewith two different isolations with similar results.
5.4. Capillary Morphogenesis in Matrigel. Matrigel(10mg/mL; BD Biosciences) was applied at 0.5mL/wellof a 6-well tissue culture dish and incubated at 37∘C for atleast 30minutes to harden. Cells were removed using trypsin-EDTA, washed with growth medium once, and resuspendedat 1 × 105 cells per mL in serum-free growth medium. Forcoculture experiments, retinal endothelial cells and pericyteswere used at a 1 : 1 ratio as we previously described [28].Cells (2mL) were gently added to the Matrigel coated plates,incubated at 37∘C, monitored at ∼18 h, and photographedusing a Nikon microscope equipped with a digital camera.In some cases, 10 𝜇g/mL of control IgG or sFlt1-FC chimera(R&D Systems) was added to the coculture experiments. Fora quantitative assessment of the data the mean number ofbranch points in 8 fields (×40) was determined. A longerincubation of the cells did not result in further branchingmorphogenesis [27].
5.5. Cell Adhesion Assays. Cell adhesion to various extracel-lular matrix proteins was performed as previously described[9]. Varying concentrations of fibronectin, human type I andIV collagen, and vitronectin (BD Biosciences) prepared inTBS with Ca2+Mg2+ (2mM each; TBS with Ca2+Mg2+) werecoated on 96-well plates (50 𝜇L per well; Nunc MaxiSorpplates, Fisher Scientific) overnight at 4∘C. Plates were thenrinsed four times with TBS with Ca2+Mg2+ and blockedwith 200𝜇L of 1% BSA prepared in TBS with Ca2+Mg2+ forat least 1 h at room temperature. Cells were removed using1.5mL dissociation solution, washed once with TBS, andresuspended at 5 × 105 cells/mL in HEPES-buffered saline(25mM HEPES, pH 7.60, 150mMNaCl, and 4mg/mL BSA).After blocking, plates were rinsed with TBS Ca2+Mg2+ once,50𝜇L of cell suspension was added to each well containing50𝜇L of TBS with Ca2+Mg2+, and the cells were allowedto adhere for 90min at 37∘C in a humidified incubator.Nonadherent cells were removed by gently washing the platefour times with 200𝜇L of TBS with Ca2+Mg2+ until no cellswere left in wells coated with BSA. The number of adherentcells in each well was quantified by measuring the levels of
intracellular acid phosphatase. Cells were lysed in 100 𝜇L oflysis buffer (50mM sodium acetate pH 5.0, 1% Triton X-100,and 4mg/mL p-nitrophenyl phosphate) and incubated at 4∘Covernight. The reaction was neutralized by adding 50 𝜇L of1M NaOH, and the absorbance was determined at 405 nmusing a microplate reader (Thermomax, Molecular Devices).All samples were prepared in triplicate, and the experimentswere repeated at least three times with similar results.
5.6. Western Blot Analysis. Cells were plated at 4 × 105in 60mm dishes coated with 1% gelatin (endothelial cells)or uncoated (pericytes) and allowed to reach nearly 90%confluence in 2 days. The cells were then rinsed oncewith serum-free medium and incubated with serum-freeDMEM for 48 hours. Then, conditioned medium (2mL)was collected and clarified by centrifugation. The 45𝜇Lof the sample was mixed with appropriate volume of 6XSDS buffer and analyzed by SDS-PAGE (4–20% Tris glycinegel; Invitrogen, Carlsbad, CA). In some cases, total proteinlysates were prepared from these cells in a modified RIPAbuffer (142.5mM KCl, 5mM MgCl
2
, 10mM HEPES, pH7.4, 2mM orthovanadate and 2mM sodium difluoride, 1%Nonidet P-40, and a complete protease inhibitor cocktail(Roche, Mannheim, Germany)). The proteins were trans-ferred to a nitrocellulose membrane, and the membrane wasincubated with an antifibronectin (Sigma), a rabbit anti-chicken tenascin-C polyclonal antibody (Thermo Scientific;Pierce, Rockford, IL, USA), anti-TSP1 monoclonal antibody(Clone A6.1, NeoMarker, Fremont, CA, USA), antiosteopon-tin (R&D, Minneapolis, MN), anti-𝛽-actin (Thermo Scien-tific; Pierce, USA), anti-HSP90 (Cell Signaling Technology),anti-Akt (Cell Signaling) antiphospho-Akt (Cell Signaling),antiphospho-eNOS (Cell Signaling), and anti-eNOS (SC-654;Santa Cruz Technology, Santa Cruz, CA). Membranes werewashed, incubated with horseradish-peroxidase-conjugatedsecondary antibody (1 : 5000, Jackson ImmunoResearch Lab-oratories, West Grove, PA) for 1 hour at room temperature,and the protein was visualized according to the chemi-luminescent procedure (Chemiluminescence reagent; GEBiosciences) [2, 14].
5.7. Flow Cytometry. Flow cytometry was performed aspreviously described [14]. The cells were washed once withPBS containing 0.04% EDTA and incubated with 2mL ofdissociation solution (Sigma) to remove the cells from theplate. The cells (106) were washed with TBS, blocked inTBS containing 1% goat serum on ice for 20 minutes, andincubated with the appropriate dilution of primary antibody,anti-PECAM-1 (BD Pharmingen), antivascular endothelial(VE)-cadherin (Alexis Biochemical, San Diego, CA), anti-𝛽1(Millipore), anti-𝛼5 (MAB1949; Millipore), anti-𝛼2 (Chemi-con), anti-𝛽3 (MAB1957; Millipore), anti-𝛼v𝛽3 (MAB1976Z;Millipore), control IgG (Jackson), rabbit anti-NG2 (AB5320;Millipore, Temecula, CA), and rat anti-mouse PDGFR𝛽(eBiosciences, San Diego, CA). For antibodies that requiredcell permeabilization, cells were removed, washed with PBS,fixed with 2% paraformaldehyde on ice for 30min, washed
14 International Journal of Cell Biology
with PBS, and resuspended in PBS containing 0.1% Triton-X-100 and 0.1% BSA containing appropriate dilution of primaryantibody.The cells were washed with TBS containing 1% BSAand then incubated with the appropriate FITC-conjugatedsecondary antibody (Jackson ImmunoResearch, West Grove,PA) for 30min on ice. After the incubation, the cells werewashed twice with TBS containing 1% BSA and resuspendedin 0.5mL of TBS containing 1% BSA. Analysis was performedon a FACScan caliber flow cytometer (Becton-Dickinson,Franklin Lakes, NJ).
5.8. Transwell Assay. Transwell filters (Corning, Acton, MA)were coated with 2𝜇g/mL fibronectin in PBS and incubatedovernight at 4∘C. The bottom of the transwell was rinsedwith PBS and blocked with 2% BSA in PBS for 1 h at roomtemperature. The transwell was rinsed with PBS, and 500𝜇Lserum-freeDMEMwas added to the bottom of eachwell, and1× 105 cells in 100 𝜇Lof serum-freemediumwere added to thetop of the transwell membrane. Following 4 hours in a 33∘Ctissue culture incubator, the cells and medium were aspiratedand the upper side of the membrane was wiped with a cottonswab. The cells that migrated through the membrane werefixed with 4% paraformaldehyde, stained with hematoxylin-eosin, and mounted on a slide. Ten-high power fields (×200)of cells were counted for each condition and the average andstandard error of the means were determined. All sampleswere prepared in duplicate, and the experiment was repeatedat least three times with similar results.
5.9. VEGF Analysis. VEGF protein levels produced by vascu-lar cells were determined using aMouse VEGF Immunoassaykit (R&DSystems,Minneapolis,MN). Cells were plated at 6×105 cells on 60mm tissue culture dishes and allowed to reachapproximately 90% confluence. The cells were then rinsedonce with serum-free DMEM and were grown in serum-freemedium for 2 days. Conditioned medium was centrifugedat 400×g for 5min to remove cell debris, and 50 𝜇L wasused in the VEGF Immunoassay. The assay was performedin triplicate as recommended by the manufacturer and wasnormalized to the number of cells. The amount of VEGF wasdetermined using a standard curve generated with knownamounts of VEGF in the same experiment. The assay wasrepeated twice using two different isolations of endothelialcells with similar results.
5.10. Nitric Oxide Analysis. Cells were plated in black wallclear bottom Microtest 96-well plates (BD #35 3948; 1 × 104cells in 100 𝜇L). The next morning the medium was changedto endothelial cell or pericyte medium containing 30𝜇MDAF-FM diacetate (Invitrogen; D-23842) and 5 𝜇g/mL ofCellTracker Red (Invitrogen; C34552). Following a 40minincubation at 33∘C, fresh mediumwas placed on the cells andthe incubation continued for 20min. The wells were washedwith TBS and the cells in each well were resuspended in100 𝜇L of TBS. Absorbance was read at 495/515 nm using afluorescence plate reader [2]. These experiments were per-formed in triplicate and repeated twice with similar results.
5.11. Statistical Analysis. Statistical differences betweengroups were evaluated with an unpaired 𝑡-test (two-tailed).Mean ± standard deviations are shown. 𝑃 values <0.05 wereconsidered significant.
Conflict of Interests
The authors declare that they have no conflict of interests.
Authors’ Contribution
Margaret Morrison, Tammy L. Palenski, and Nasim Jamaliperformed experiments and analyzed data. Margaret Morri-son, Tammy L. Palenski, Nasim Jamali, Nader Sheibani, andChristine Sorenson conceived and designed the experiments.Christine Sorenson and Nader Sheibani prepared the paper.
Acknowledgments
This work was supported by Grants from the University ofWisconsin Department of Pediatrics Research and Develop-ment Fund (Christine M. Sorenson). Christine M. Sorensonwas funded, in part, by EY 023024 from the NationalInstitutes of Health and the Retina Research Foundation.Nader Sheibani is supported by NIH Grants EY016995,EY018179, and RC4 EY 021357, P30 CA014520 UW PaulP. Carbone Cancer Center Support Grant, P30 EY016665,and an Unrestricted Departmental Award from Researchto Prevent Blindness. Nader Sheibani is a recipient of aResearch Award from American Diabetes Association (1-10-BS-160) and Retina Research Foundation. Tammy L. Palenskiis a recipient of a Kirschstein-NRSA Fellowship Award(F31 EY021091) and NIH Grant T32 ES007015. Margaret E.Morrison is a recipient of a Senior Thesis Grant from theCollege of Letters and Science at the University ofWisconsin-Madison. The authors wish to thank Robert Gordon for hisassistancewith the graphics and SunYoungPark for theVEGFanalysis.
References
[1] R. J. Youle andA. Strasser, “The BCL-2 protein family: opposingactivities thatmediate cell death,”Nature ReviewsMolecular CellBiology, vol. 9, no. 1, pp. 47–59, 2008.
[2] C. Grutzmacher, S. Park, T. L. Elmergreen et al., “Opposingeffects of bim and bcl-2 on lung endothelial cell migration,”TheAmerican Journal of Physiology—Lung Cellular and MolecularPhysiology, vol. 299, no. 5, pp. L607–L620, 2010.
[3] P. Bouillet, M. Robati, M. Bath, and A. Strasser, “Polycystickidney disease prevented by transgenic RNA interference,” CellDeath and Differentiation, vol. 12, no. 7, pp. 831–833, 2005.
[4] P. Bouillet, S. Cory, L. C. Zhang, A. Strasser, and J. M. Adams,“Degenerative disorders caused by Bcl-2 deficiency preventedby loss of its BH3-only antagonist bim,”Developmental Cell, vol.1, no. 5, pp. 645–653, 2001.
[5] P. Bouillet, D. Metcalf, D. C. Huang et al., “Proapoptotic Bcl-2relative Bim required for certain apoptotic responses, leukocytehomeostasis, and to preclude autoimmunity,” Science, vol. 286,no. 5445, pp. 1735–1738, 1999.
International Journal of Cell Biology 15
[6] R. Ley, K. E. Ewings, K. Hadfield, and S. J. Cook, “Regulatoryphosphorylation of Bim: sorting out the ERK from the JNK,”Cell Death andDifferentiation, vol. 12, no. 8, pp. 1008–1014, 2005.
[7] M. Fruttiger, “Development of the retinal vasculature,” Angio-genesis, vol. 10, no. 2, pp. 77–88, 2007.
[8] S. Wang, Z. Wu, C. M. Sorenson, J. Lawler, and N. Sheibani,“Thrombospondin-1-deficient mice exhibit increased vasculardensity during retinal vascular development and are less sensi-tive to hyperoxia-mediated vessel obliteration,” DevelopmentalDynamics, vol. 228, no. 4, pp. 630–642, 2003.
[9] M. J. Pollman, L. Naumovski, and G. H. Gibbons, “Endothelialcell apoptosis in capillary network remodeling,” Journal ofCellular Physiology, vol. 178, no. 3, pp. 359–370, 1999.
[10] K. Walsh, R. C. Smith, and H. S. Kim, “Vascular cell apoptosisin remodeling, restenosis, and plaque rupture,” CirculationResearch, vol. 87, no. 3, pp. 184–188, 2000.
[11] S.Wang, S. Park, P. Fei, and C.M. Sorenson, “Bim is responsiblefor the inherent sensitivity of the developing retinal vasculatureto hyperoxia,” Developmental Biology, vol. 349, no. 2, pp. 296–309, 2011.
[12] P. Carmeliet, “Angiogenesis in life, disease and medicine,”Nature, vol. 438, no. 7070, pp. 932–936, 2005.
[13] J. Andrae, R. Gallini, and C. Betsholtz, “Role of platelet-derived growth factors in physiology and medicine,” Genes andDevelopment, vol. 22, no. 10, pp. 1276–1312, 2008.
[14] S. Kondo, E. A. Scheef, N. Sheibani, and C. M. Sorenson,“PECAM-1 isoform-specific regulation of kidney endothelialcell migration and capillary morphogenesis,” The AmericanJournal of Physiology—Cell Physiology, vol. 292, no. 6, pp.C2070–C2083, 2007.
[15] N. Sheibani, E. A. Scheef, T. A. Dimaio, Y. Wang, S. Kondo, andC. M. Sorenson, “Bcl-2 expression modulates cell adhesion andmigration promoting branching of ureteric bud cells,” Journalof Cellular Physiology, vol. 210, no. 3, pp. 616–625, 2007.
[16] S. Kondo, Y. Tang, E. A. Scheef, N. Sheibani, and C. M.Sorenson, “Attenuation of retinal endothelial cell migration andcapillary morphogenesis in the absence of bcl-2,”The AmericanJournal of Physiology—Cell Physiology, vol. 294, no. 6, pp. C1521–C1530, 2008.
[17] J. Ziehr, N. Sheibani, and C. M. Sorenson, “Alterations incell-adhesive and migratory properties of proximal tubule andcollecting duct cells from bcl-2 -/- mice,”The American Journalof Physiology—Renal Physiology, vol. 287, no. 6, pp. F1154–F1163,2004.
[18] A. Jozkowicz, J. Dulak, A. Nigisch et al., “Involvement of nitricoxide in angiogenic activities of vascular endothelial growthfactor isoforms,” Growth Factors, vol. 22, no. 1, pp. 19–28, 2004.
[19] D. Fukumura, T. Gohongi, A. Kadambi et al., “Predominantrole of endothelial nitric oxide synthase in vascular endothelialgrowth factor-induced angiogenesis and vascular permeability,”Proceedings of the National Academy of Sciences of the UnitedStates of America, vol. 98, no. 5, pp. 2604–2609, 2001.
[20] M. G. Blanes, M. Oubaha, Y. Rautureau, and J. Gratton,“Phosphorylation of tyrosine 801 of vascular endothelial growthfactor receptor-2 is necessary for Akt-dependent endothelialnitric-oxide synthase activation and nitric oxide release fromendothelial cells,” Journal of Biological Chemistry, vol. 282, no.14, pp. 10660–10669, 2007.
[21] E. A. Scheef, C. M. Sorenson, and N. Sheibani, “Attenuation ofproliferation andmigration of retinal pericytes in the absence ofthrombospondin-1,” The American Journal of Physiology—CellPhysiology, vol. 296, no. 4, pp. C724–C734, 2009.
[22] M. Saint-Geniez and P. A. D’Amore, “Development and pathol-ogy of the hyaloid, choroidal and retinal vasculature,” Interna-tional Journal of Developmental Biology, vol. 48, no. 8-9, pp.1045–1058, 2004.
[23] M. R. Hayden, Y. Yang, J. Habibi, S. V. Bagree, and J. R. Sowers,“Pericytopathy: oxidative stress and impaired cellular longevityin the pancreas and skeletal muscle in metabolic syndrome andtype 2 diabetes,”Oxidative Medicine and Cellular Longevity, vol.3, no. 5, pp. 290–303, 2010.
[24] S. Wang, C. M. Sorenson, and N. Sheibani, “Attenuation ofretinal vascular development and neovascularization duringoxygen-induced ischemic retinopathy in Bcl-2-/- mice,” Devel-opmental Biology, vol. 279, no. 1, pp. 205–219, 2005.
[25] J. I. Greenberg, D. J. Shields, S. G. Barillas et al., “A role forVEGF as a negative regulator of pericyte function and vesselmaturation,” Nature, vol. 456, no. 7223, pp. 809–813, 2008.
[26] C. Grutzmacher, S. Park, Y. Zhao, M. E. Morrison, N. Sheibani,and C. M. Sorenson, “Aberrant production of extracellularmatrix proteins and dysfunction in kidney endothelial cellswith a short duration of diabetes,” The American Journal ofPhysiology—Renal Physiology, vol. 304, no. 1, pp. F19–F30, 2013.
[27] X. Su, C. M. Sorenson, and N. Sheibani, “Isolation and charac-terization of murine retinal endothelial cells,”Molecular Vision,vol. 9, pp. 171–178, 2003.
[28] T. L. Palenski, C. M. Sorenson, C. R. Jefcoate, and N. Sheibani,“Lack of Cyp1b1 promotes the proliferative and migratoryphenotype of perivascular supporting cells,” Laboratory Inves-tigation, vol. 93, no. 6, pp. 646–662, 2013.